Picking the Plough or Not

How do personality traits influence occupational attainment in emerging

Southeast Asia?∗

Dorothee Buhler† Rasadhika Sharma‡ Wiebke Stein§

26th August 2019

WORK IN PROGRESS. PLEASE DO NOT CITE OR CIRCULATE WITHOUT PERMISSION.

Abstract

Rural labor markets in emerging countries offer limited occupation choices to individuals and decisions

are often made out of necessity rather than opportunities. Given these difficult circumstances it is

all the more important to understand the role of personality traits for occupation related decisions.

Against this background, this paper examines the relation between personality traits and occupational

attainment in two emerging Southeast Asian economies. Using new micro level data from individuals

living in rural areas of Thailand and Vietnam we test the importance of a broad range of personality

measures, while controlling for other channels of human capital formation. In addition, we are able to

contrast the influence of personality traits on occupational attainment for both an emerging market

and a socialist economy. Our results suggest that common measures of personality traits can be

applied in a rural setting and are important for individual occupational attainment. In particular, we

find that conscientious individuals are more likely to engage in self-employment or permanent jobs.

Other non-cognitive skills such as locus of control, trust, risk, and patience play an inferior role.

Keywords: Personality traits; Big Five Factor Model; Occupational Attainment; Southeast Asia;

TVSEP

JEL: D91; O1; R2

∗We thank Ulrike Grote, Stephan Thomsen, Andreas Wagener, and seminar participants in Gottingen and Hannover fortheir helpful comments. Financial support by the German Research Foundation (DFG) via the Research Training Group1723 and the Thailand Vietnam Socio Economic Panel is gratefully acknowledged.

†Leibniz University of Hanover, Germany; Email: buehler@iuw.uni-hannover.de‡Leibniz University of Hanover, Germany; Email: sharma@iuw.uni-hannover.de§Leibniz University of Hanover, Germany; Email: stein@wipol.uni-hannover.de

1 Introduction

Determining individual’s occupation decisions is a prominent topic among labor economists. Empirical

and experimental evidence (Jencks and Williams, 1979; Nyhus and Pons, 2005; Wells et al., 2016) cor-

roborate that in addition to cognitive skills, non-cognitive skills – also called personality traits – play an

important role in determining individual decision making behavior (Wichert and Pohlmeier, 2010), job

performance (Barrick and Mount, 1991) and economic outcomes (Piatek and Pinger, 2010). According to

Heckman and Cunha (2007) non-cognitive skills are vital when it comes to actually realizing the acquired

potential (i.e. cognitive skills). Therefore, the importance of personality traits in determining occu-

pational attainment (Cobb-Clark and Tan, 2011; John and Thomsen, 2014) and occupational earnings

(Osborne Groves, 2003; Mueller and Plug, 2006; Heineck and Anger, 2010) has been acknowledged widely

in the literature. While existing studies have produced valuable results, the focus is on industrialized

countries, using data from the U.S., Japan or European countries.

By 2030, 85 % of the worlds population will be based in emerging and developing countries, building

the backbone of the global economy (UNHCS, 2001). In order to create sufficient job opportunities for the

growing population and improve living circumstances, it is vital to understand individual labor market

decisions in these countries. Labor markets in emerging and developing countries vary substantially

from labor markets in industrialized countries. They are characterized as labor intensive and capital

scarce (Campbell and Ahmed, 2012). There is little specialization and mostly small scale operations

(Banerjee and Duflo, 2007). In particular, rural economies are informal, likely to be less productive, credit

insufficient, and prone to greater earnings instability (Campbell, 2011). Gollin et al. (2014) highlight

that in most developing countries labor is misallocated into the agricultural sector, leading to a large

agricultural productivity gap. Low levels of skills among farmers is thought to be one of the main reasons

for this gap (Laajaj and Macours, 2017). Therefore, transferring results from existing studies would be

over-simplification. Only recently, some studies started to address the determinants and impacts of non-

cognitive skills in developing economies (Attanasio et al., 2015; Gertler et al., 2014; Laajaj and Macours,

2017).

Against this background, this paper aims to shed light on personality traits and their relevance in

rural labor markets of emerging Southeast Asian economies. The scope of this study is twofold: First,

we implement and validate measures of personality traits, namely the Big Five factor model (McCrae

and John, 1992; Costa and McCrae, 1997), in an emerging country setting. Second, we examine the

importance of personality traits for occupational attainment in rural areas of Southeast Asia. Given that

personality traits influence preferences and individual’s motivation, it is worth understanding why some

individuals decide to pursue an occupation other than farming as their main occupation. Since the skill

set and the labor market opportunities in these settings are rather homogeneous, personality traits might

explain why some individuals pursue a career other than farming. We therefore employ a multinominal

logit model, using subsistence farming as the baseline occupation and test it against different occupation

categories such as self-employment or permanent wage employment. In addition, we are able to study

the importance of personality traits in two different political systems.

For our analysis, we use a data set for Thailand and Vietnam, collected under the Thailand Vietnam

Socio Economic Panel (TVSEP) in 2017.1 A section on measurement of personality traits was included

for the first time in 2017, providing information on around 4000 individuals. In our sample, like most

rural populations in developing countries, the majority of households rely on income from agriculture

and environmental resource extraction (Parvathi and Nguyen, 2018). Previous research in the region

1 For more information please refer to the project webpage: https://www.tvsep.de/overview-tvsep.html.

1

shows that households engaging in activities other than farming tend to be better off than farming-only

households (Sohns and Revilla Diez, 2016; Sharma et al., 2016).

In order to capture an individual’s personality we use eight different measures: the Big Five factor

model, locus of control as well as measures on patience, risk and trust. The Big Five model thereby

captures a persons intrinsic personality, whereas the other measures determine outcome related behavior

and attitude (John and Thomsen, 2014).

Our results confirm that the Big Five personality measures hold for our rural sample population.

Further, our robustness tests show that the measures are stable across time. Despite the different political

regimes, the results do not differ substantially between the Thai and Vietnamese sub sample. In terms

of occupational attainment, we find that Conscientiousness is the most decisive factor. This is line

with literature utilizing developed country data sets (Barrick and Mount, 1991). Conscientiousness is

particularly important for choosing high skilled jobs over subsistence farming. Our results indicate that it

is crucial for development support policies to identify individuals who are willing to choose an occupation

outside of the farming business. Our study adds to the literature on personality measures by applying

existing measures to individuals in rural Southeast Asia. In addition, we add to the literature on the

importance of personality traits for occupational attainment by providing evidence from a large panel

study in Thailand and Vietnam.

The rest of the paper is organized as follows: Section 2 provides the theoretical framework for our

paper. Section 3 introduces the study design and illustrates data collection, measurement of the traits

and the econometric models used in our paper. Section 4 presents the results, which is followed by a

conclusion in Section 5.

2 Theoretical Framework

2.1 Occupation and human capital

Labor market outcomes, such as occupational attainment and earnings, relate to the human capital of

the individual. To conceptualize this, we follow the human capital and earnings models that study the

behavior of humans in relation to their human capital formation and occupational attainment (Ben-

Porath, 1967, 1970; Mincer, 1970; Heckman, 1976; Cunha et al., 2006). In the simplest setting, we

assume that individuals choose their occupation in order to maximize their life-time earnings. Further,

individuals are assumed to choose the individually optimal investment into human capital that allows

them to reach their desired occupation and to maximize their earnings (Blau et al., 1956). Thus, labor

market outcomes, such as occupational attainment and earnings, can be depicted as:

max(LMoutcome) = f(H) + ε (1)

Where ε signifies the idiosyncratic difference in labor market outcomes and H the human capital of

the individual. Human capital is considered a latent variable as there does not exist one variable that

allows us to measure it directly. It is rather a combination of different factors which define it, including

skills (S) and other individual characteristics (I), socio-economic characteristics related to the family

background (F), and other factors, which affect the labor market outcome (X), including labor market

experience, labor market conditions and health (Mincer, 1970, 1974). Therefore, human capital can be

2

formally described as:

H = αS + βI + δF + γX + µ (2)

The individual skill set S refers to cognitive and non-cognitive skills, which are key determinants

for labor market outcomes of the individual. Cognitive skills, often referred to as education or level of

technical skills, are closely associated with the individuals occupational attainment as well as earning

differentials (Cawley et al., 2001; Finnie and Meng, 2002; Hanushek and Woessmann, 2008; McIntosh

and Vignoles, 2001). In addition to cognitive skills, non-cognitive skills, such as personality traits or

interpersonal skills, relate to labor market outcomes. In particular, empirical evidence suggests that

personality traits affect job search behavior, occupational attainment, job satisfaction, work behavior,

and income (Ones et al., 2003; Judge et al., 2002b,a).

In addition to cognitive and non-cognitive skills other characteristics related to the individual (I) such

as gender, age and ability shape the occupation decision and the earnings potential. In the context of

developing countries gender plays an important role with respect to labor market outcomes as females face

different constraints compared to men. Evidence from Schmidt and Strauss (1975) suggests that females

are less likely to work in high skilled white collar jobs. Besides those individual characteristics, the

socio-economic and family background matters for the formation of human capital and thus labor market

related decisions (F ). Social capital is also an important factor. Evidence from (Bentolila et al., 2010)

shows that the the network of the individual or family increase the chance of finding job opportunities.

Finally, the general labor market situation and other regional disparities are decisive for the individual

decision making and realization from human capital investments (X).

We therefore capture these aspects in our model and include measures for each facet of human capital

into the analysis.

2.2 Definition of non-cognitive skills

As was previously described, personality traits build a part of the individuals human capital. Besides the

Big Five model, which is the most prominent concept, there are other measures that capture a persons

personality, such as locus of control, trust, risk attitudes, and patience.

The Big Five Inventory The Big Five model proposed by McCrae and John (1992); Costa and Mc-

Crae (1997) is the most cross-culturally validated model of personality traits (Stuetzer et al., 2018). The

factors are relatively stable over an individual’s lifetime (Heineck and Anger, 2010) and are considered

heritable (Hofstede and McCrae, 2004). The Big Five model includes traits of Openness, Conscien-

tiousness, Extraversion, Agreeableness, and Neuroticism.2 Openness captures how individuals value new

experiences and changes (Rolland, 2002). An open person is creative and enthusiastic about complex

jobs. Previous research finds that individuals who are more open, opt for self-employment (Obschonka

and Stuetzer, 2017; Stuetzer et al., 2018) or prefer professional jobs requiring analytical and creative

thinking (Wells et al., 2016; John and Thomsen, 2014). Conscientiousness depicts how an individual han-

dles tasks. Persons displaying high levels of Conscientiousness are responsible, efficient and hardworking,

in their own work and the work done for others (Wichert and Pohlmeier, 2010). Extraversion captures the

individual’s social relationship. A person with a high level of Extraversion seeks to establish contact with

others, displays confidence and is positive (Schafer, 2016; Wichert and Pohlmeier, 2010). Extroverted

2 See Table A.1 in the appendix for a graphic overview

3

individuals are expected to choose and perform better in jobs involving social interaction (Barrick and

Mount, 1991). Agreeableness refers to the quality of interpersonal relationships of the individual. An

agreeable person is caring and selfless. Neuroticism captures how an individual behaves under stressful

situations. Scoring high on this factor indicates that the individual is emotionally unstable and does not

cope well with stress (Rammstedt, 2007; Schmitt et al., 2008). Overall, Conscientiousness is considered

as the most important predictor of occupational performance (Barrick and Mount, 1991). In case of

Openness, there is no consensus on the influence of higher Openness on earnings, with studies demon-

strating both positive (Mueller and Plug, 2006) and negative association (Seibert and Kraimer, 2001).

An extraverted person earns more and is more successful at work, while an agreeable individual would

display lower job satisfaction (Seibert and Kraimer, 2001). Agreeableness is also linked positively to the

individual’s normative commitment (Erdheim et al., 2006). Scoring high on Neuroticism negatively influ-

ences earnings and job satisfaction (Nyhus and Pons, 2005). Specifically for entrepreneurship, a successful

entrepreneur scores high on Extraversion, Conscientiousness, and Openness and low on Agreeableness

and Neuroticism (Stuetzer et al., 2018; Obschonka and Stuetzer, 2017).

Additional non-cognitive skills Another non-cognitive skill is the locus of control, which captures

the individuals belief of how much their decisions affect their outcomes (Rotter, 1966). A person with

an internal locus of control believes that reinforcement in life is contingent on their actions (Piatek and

Pinger, 2016). In contrast, a person with an external locus of control views their life as being beyond

their control and, rather, influenced by external factors such as destiny (Caliendo et al., 2015). We

expect that individuals with an internal locus of control to be more likely to step out of their comfort

zone. Hence, this trait might be stronger in entrepreneurs or managers and less visible in professionals

(John and Thomsen, 2014). Additionally, it is proposed that individuals with a stronger internal locus

tend to invest in themselves, for example in education and training (Piatek and Pinger, 2010) and in

their businesses and employees (Sharma and Tarp, 2018). These individuals also engage in high paying

jobs and show greater mobility towards higher paying jobs (Schnitzlein and Stephani, 2016). Trust is the

quality of an individual to rely on others, to trust them and their dealing of strangers (Caliendo et al.,

2012). Caliendo et al. (2014) find that a trustworthy person is more likely to start a business. The most

researched personality trait in labor force participation is the risk attitude of the individual (Caliendo

et al., 2010). If the individual is risk averse, they will prefer more stable job profiles such as permanent

employment or the default of subsistence farming. On the contrary, a risk-loving individual is more

likely to engage in self-employment or adopt new technologies (Dustmann et al., 2017; Falk et al., 2018).

Another variable is patience, which refers to the individuals willingness to wait. Patient people are more

likely to save and disply higher educational attainment (Falk et al., 2018) which might lead to better

occupational outcomes. Literature posits that being patient might positively influence entrepreneurial

decisions (Caliendo et al., 2012). However, Caliendo et al. (2014), find that patience is collinear with

emotional stability or Neuroticism included in the Big Five, and, therefore, would display no effect on

the occupational decision. We are unclear about the expectations in this regard.

3 Study Design

3.1 Country Background and Data

We use micro data originating from the Thailand Vietnam Socio Economic Panel (TVSEP). The eco-

nomic, cultural and institutional background across the two countries is quite diverse. While Thailand is

4

a constitutional monarchy which operates under relatively free, market driven policies, Vietnam belongs

to the four remaining countries worldwide which are governed by a one-party socialist system openly

advocating communism (Gloede et al., 2015).3 Economic development across Thailand and Vietnam

differs substantially. According to their Gross Domestic Product (GDP) per capita, Vietnam belongs

to the lower-middle-income economies, while Thailand is classified as an upper-middle income country

(World Bank, 2018). Despite recent growth and increases in overall household wealth, pockets of poverty

persist in rural areas of both countries (Hardeweg et al., 2013b).

Since 2007, the TVSEP regularly administers surveys among rural households in Thailand and Viet-

nam. Until now, six additional waves have been conducted, in 2008, 2010, 2011, 2013, 2016 and 2017.

The Thai data were collected in the provinces Buriram, Nakhon Panom and Ubon Ratchathani and the

Vietnamese data in the provinces Thua Thien Hue, Ha Tinh and Dak Lak. Figure A.1 in the Appendix

exhibits an overview of the survey region. The survey covers 4,000 rural households in 440 villages. For

the purpose of this study, we use data on 2,734 individual respondents who answered the subsection on

personality traits. The sample is not exactly identical to the household sample due to three reasons:

First, common survey attrition; Second, we have to exclude households that did not answer the survey

items; Third, we apply an age restriction and only include respondents aged between 17 to 64 years

because our analysis focuses on working-age individuals (OECD, 2019).

The household sample in each province was randomly drawn based on a stratification process con-

sidering the heterogeneous agro-ecological conditions within the regions.4 All monetary variables were

converted to 2005 Purchasing Power Parity USD (PPP USD) equivalents.

In both countries, an almost identical household survey is applied. It consists of nine sections covering

individual information on household members (e.g. age, education, health, and employment) as well as

household-level information on expenditures, shocks, risks, income earning activities such as farming,

livestock raising and fishing, household financial situation, housing conditions, transfers received, and

assets owned. In addition to the household survey, a village-level survey is administered to the village

chief collecting information on the village location, population, infrastructure, employment, agriculture,

and economic conditions.

In the 2017 panel wave of the TVSEP, an additional module was included which asks for the estab-

lished psychological personality inventories. These questions allow to study personality traits and their

consequences on a large, representative sample of rural households in Thailand and Vietnam and to relate

them to a rich set of socio-economic variables.

For our analysis, we categorize our sample of working age individuals by occupation type and define

seven occupation categories: Subsistence farming, commercial farming, self-employed, permanent wage

employed, worker (casual), housewife and not working.5 We separate between subsistence farming and

commercial farming to get a more comprehensive understanding of peoples job opportunities.6 The survey

setting is rural areas in Thailand and Vietnam. Naturally, people mainly work in farming. In our sample,

67% of the respondents report agriculture as their main occupation, of those 67%, 42% are engaged in

subsistence farming, meaning that their agricultural output is mainly used for home consumption and not

for commercial gain. 25% are engaged in commercial farming. The remaining 33% of the sample engage

in different wage generating activities. The majority of people that do not work in the farming sector

are self-employed (12%). Respondents run various kinds of businesses, for example retail or small food

3 The other three countries are China, Cuba and Lao PDR.4 See Hardeweg et al. (2013b) for a detailed overview of the sampling strategy.5 We build the occupation categories based on self reported answers from the respondent about their main occupation.6 Respondents are categorized as being engaged in subsistence farming, if the share of income from farming of the respon-

dents household is less than 0.5. If the farming share from the household income is larger than 0.5, the respondent issaid to be engaged in commercial farming.

5

shops. 9% are employed permanently. Respondents engaged in permanent wage employment have on

the highest education, with an average of 10 years. 6% of respondents are casual workers, 4% housewives

and 2% do not work.

In order to validate our survey measures we use additional data from a TVSEP Add-on project that

was conducted in Ubon Rathathani, Thailand, in November 2017 amongst the same households. 7 The

Add-on questionnaire includes the exact same questions on personality traits as the TVSEP household

survey from summer 2017. This gives us the unique opportunity to compare the answers from one

individual at two different points in time for one of the TVSEP survey provinces. Hence, we can test

the stability of the survey measure. We identified 505 cases where the respondent in the summer and in

November are the same person. For these 505 cases, we compare the answers given in the summer with

those given in November.

3.2 Measurement of personality traits

We capture the different aspects of personality by using nine distinct measures: the Big Five inventory,

locus of control, risk and trust. Apart from the Big Five Inventory, all items have been asked and tested

in previous survey waves.

The Big Five Inventory We follow the Big Five model (Costa and McCrae, 1992, 1997) which has

become the standard personality measurement in psychology, described in Section 2.2. The model defines

personality along the five following dimensions: Openness, Conscientiousness, Extraversion, Agreeable-

ness, and Neuroticism. The survey questions used to capture these dimensions are based on the Big Five

personality inventory questions used in the German Socio Economic Panel (SOEP).8 Similar questions

are used in the British micro panel survey and World Bank surveys across different countries (Guerra et

al., 2016). In the respective questionnaire section in the TVSEP survey, respondents are asked how much

they agree with different statements about themselves. They rank their answers on a 7 point Likert scale

ranging from 1 to 7, where 1 means ”Does not apply to me at all” and 7 means ”Applies to me perfectly”.

In total, respondents are presented with 15 survey questions. Figure A.2 in the Appendix exhibits an

overview of the survey questions.9

Locus of control We capture the extent to which respondents believe that they have control over

their life outcomes by using a survey item that asks about the reasons for why people have low incomes.

Answers include among others: pure luck, knowing the right people, or hard work good education. The

item scale ranges from 1 to 9, where 1 means that the person has a complete external locus of control

and 9 means that the person has a complete internal locus of control.

Risk We capture someones willingness to take risks by using the standard risk measurement item

used in many economic studies. The question asks: Are you generally a person who is fully prepared to

take risks or do you try to avoid taking risk? Respondents can rank themselves on a scale from 0 to 10.

7 The Add-on project is about Behavioral insights into over-indebtedness within a vulnerable population. For more detailson the Add-on project, see Kluhs et al. (2019).

8 See survey page for details: https://www.diw.de/en/soep9 An identical set of questions was administered to individuals who participated in the Add-on project. However, the answer

modalities differed slightly. Although, the items are measured on the same scale (7 point Likert scale), each number onthe scale was labelled explicitly (each answer option is associated with a specific phrase, e.g. 1 means Disagree fully, 3means Disagree a little, 6 means Agree strongly). Despite these differences, we rely on the comparison of the TVSEPdata with the Add-on data to reveal, if the measures are reliable or not.

6

0 means Unwilling to take risks and 10 means Fully prepared to take risk. This survey measure has been

experimentally validated for the TVSEP data by Hardeweg et al. (2013a)

Trust Trust is measured by a dummy variable, based on the following question: Generally speak-

ing, would you say that most people can be trusted or that you need to be very careful when dealing with

people? The variable is 0 if the respondents answer is Need to be very careful when dealing with people

and 1 if the answer is Most people can be trusted. Similar scales have been applied in the SOEP and

validated by Becker et al. (2012); Dohmen et al. (2009).

Patience We measure a persons patience through the following question: Are you generally a

person who is fully prepared to give up something now in order to gain more in the future? Again,

respondents rank themselves on a scale from 0 to 10, where 0 means Unwilling to wait and 10 means

Fully prepared to wait. Similar questions have been used in other major surveys such as the Global

Preference Survey and validated by (Falk et al., 2018, 2016).

3.3 Labor market status and characteristics of sample population

In this section we give an overview of characteristics of the whole TVSEP population. In total, the survey

population covers information on about 11,000 working-age individuals. Working-age individuals refers to

people aged 17-64 (OECD, 2019). While we have individual information about all working-age household

members, the personality traits section was only answered by the household head or his/her spouse.

However, we use information on the total sample of working-age individuals to shed light on the overall

labor market situation in our study areas. Hereby, we want to give an idea about the differences between

our sample population – that consists only of the respondents – and the rest of the survey population,

which is representative for rural households in these areas. Table 1 displays the main characteristics of

the working-age population and the respondents and Figure 1 shows the employment status and sector

composition of employment for the working-age population10 in our sample.

The average working-age individual in our survey is about 38 years old, has 8.6 years of education and

belongs to the ethnic majority of the respective country (Thai or Kinh). The gender ratio is on par and

every second person reports to be engaged in farming. The distribution of employment status shows that

a large share (40 %) of the working-age population is engaged in either subsistence or commercial farming.

Given that many households operate their own farming business the labor market status of these people

is similar to being self-employed. About one fourth of the working-age individuals are permanently

employed, 9 percent are in casual employment, and 8 percent are self-employed outside agriculture.

While agriculture is clearly the most important sector for employment, production & industry and crafts

& services together account for one fourth of the employment.

As Table 1 shows the respondents differ from the rest of the working-age population in relation to age,

gender, literacy and education. Given that the project aims to interview the household head or his/her

spouse this is not surprising. However, it also means that our results can not be generalized to the total

working-age population in the survey areas. The average respondent in our sample is engaged in farming,

49 years old, female, and has 6 years of education.

10 We us data on all working-age individuals collected in the survey. Working-age refers to individuals aged 17 to 64 yearsof age

7

Table 1: Descriptive statistics of working-age sample population

Thailand VietnamSample Respon- Rest of Sign. Respon- Rest of Sign.working-age dents working-age dents working-agepopulation population population

Age (years) 37.79 47.74 35.82 *** 50.51 32.61 ***Gender (in %) (1=Female) 0.50 0.59 0.46 *** 0.70 0.46 ***Literacy (in %) 0.96 0.92 0.99 *** 0.98 0.94 ***Education (years) 8.63 6.95 9.58 *** 6.34 8.93 ***Ethnic majority (in %)* 0.87 0.76 0.97 *** 0.96 0.78 ***Religious (in %)* 0.64 0.29 1.00 *** 1.00 0.29 ***Farmer (in %) 0.50 0.84 0.38 *** 0.77 0.44 ***

N 11,292 1,643 4,287 1,540 3,822

Note: Whole sample refers to the total working-age population covered in the sample. Working-age refers to individualsaged 17 to 64. Respondents are those individuals who answered the questionnaire i.e. those individuals who respondedto the character traits and risk-related questions. *, **, and *** denote significance at the 10, 5, and 1 percent levelsfor two-sided ttests.*Shares for Religious and Ethnicity are based on answers from the respondent about the respective household memberSource: Authors’ calculations.

Figure 1: Employment status of working-age population*

(a) Occupational attainment

13.33%

40.40%

8.29%

9.11%

25.61%

3.26%

Not working Farmer Self-employedWorker (casual) Worker (permanent) Housewife

(b) Sector of employment

38.30%

13.07%7.16%4.10%

11.27%

5.51%

4.79%

3.15%

12.65%

Agriculture Production & Industry Trade, Transport & CommunicationHotel & Food Crafts & Services PublicOther Domestic Care Not working

Note: Working-age population refers to all individuals covered in the survey aged between 17 and 64 years old.Source: Authors’ calculations.

8

3.4 Specification of econometric models

In the analysis, we use different methodologies to address our research questions. We start by assessing the

internal validity of the survey measure for personality traits in our sample and use descriptive statistics to

analyze differences between occupation groups. We then apply a multinomial logit regression to estimate

the influence of personality traits on an individual’s occupational attainment.

In the first step, we address the internal validity of the Big Five model for our sample population in

three ways: (i) we compute the Cronbach’s itemized alpha coefficient to test for internal consistency of

the scales, (ii) we conduct a PCA based on the survey questions, and (iii) we test the stability of the

personality traits over time. The Cronbach’s itemized alpha coefficient (Cronbach, 1951) is widely used

in the psychological literature and tests the internal consistency of the scales across the survey questions

and across the five personality traits (Schafer, 2016; Yomaboot and Cooper, 2016).

We use a principal component analysis (PCA) to validate the structure of the personality factors for the

sample population (Schmitt et al., 2007). The PCA is based on the 15 personality questions administered

to respondents in the household questionnaire (see Table A.2) and allows it us to reduce the dimension

of the traits variables by creating factors which are homogeneous within themselves and heterogeneous

between each other (Backhaus et al., 2011). In order to compare our measures with other studies, we

also construct simple averages for the respective Big Five traits to produce comparable measures of the

personality traits for our sample population (see Tables A.1 and A.2 for relation between personality

traits and survey questions). We use the Kaiser criterion (K1) (Ford et al., 1986) which retains all factors

with eigenvalues greater or equal to one, to determine the number of factors to be retained, resulting in

five factors which explain a total of 56% of the variance. Following Hair et al. (2009), only the factors

with loadings greater than 0.30, i.e. meeting the minimum practical significance level, are interpreted.

To validate the stability of the personality traits in our sample, we use data from an Add-on project

conducted in Ubon Ratchathani. Given that the same individuals were asked the same type of questions,

this allows us to compare the responses from the same person at two different points in time. A two-sided

ttest is executed to compare the results.

In the second step of the analysis, we group individuals into seven occupation groups: (i) subsistence

farming, (ii) commercial farming, (iii) self-employed, (iv) worker casual, (v) worker permanent, (vi)

housewife, and (vii) not working. The grouping is based on the self-reported information regarding the

’main occupation in the reference period’. We use two-sided ttests to detect differences between different

types of occupations with respect to personality traits, cognitive skills, and individual characteristics such

as age, religion, ethnicity and gender.

In the thrid step, we use a multinomial logit model to understand, in how far the individual’s per-

sonality traits predict occupational attainment. The concept of the human capital approach, described

in Section 2.2 is the basis for our regression model and the variables included. The regression takes the

following form:

Pr(Oijr = 0) = β0 + β1Eijrc + +β2Pijrc + γ1Iijr + γ2HHjr + γ3LMr + γ4Dr + εijr (3)

Where Oijr denotes probability of individuals i in household j and region r to engage in an occupation

in relation to the base category of being a subsistence farmer. Eijrc represents the cognitive skills of

individuals measured as years of schooling. The vector Pijrc captures the set non-cognitive skills of each

individual, including the Big Five measures as well as the variables on locus of control, patience, risk

and trust. Iijr and Hjr are vectors of individual and household control variables. These include gender,

9

age, and being active in a political party. Furthermore we also control for household size, ethnicity and

mobile phone ownership refer to socio-economic characteristics related to the family background. Finally,

we include controls for the labor market situation (LMr) at the district level and for the province (Dr)

in which the household resided to capture regional differences.

4 Results

4.1 Validity of the survey measure

This section addresses the internal validity of our newly introduced survey measures - the Big Five Inven-

tory. We use an unrestricted sample throughout Section 4.1 and only apply the working age restriction

in Sub-section 4.2 and 4.3, where we analyze the relationship between personality traits and occupational

attainment.

Cronbach’s Alpha In order to test for internal consistency of the survey measures we compute the

Cronbach’s itemized alpha coefficient for the overall TVSEP sample as well as for the Ubon Sub-sample.

The Cronbach’s itemized alpha coefficient ranges between 0.42 and 0.60 across the Big Five factors. The

overall reliability lays a 0.64 for the whole sample indicating a good fit (Schafer, 2016) . For the Ubon

Sub-sample the score lays at 0.67. Detailed results are reported in Table A.3 in the Appendix. Our

results are similar to those of Rammstedt and John (2007).

Principal Component Analysis The PCA reveals five factors (see Table A.3 and Figure A.2 in the

Appendix). In order to avoid confusion with the five factors from the Big Five model we name our factors:

(i) Creativeness, (ii) Diligence, (iii) Skepticism, (iv) Approachability, and (v) Amiableness. Individuals

who are creative consider themselves as artistic, have new ideas and an active imagination. They work

thoroughly and efficiently, are sociable, and kind to others. People who are diligent are very determined

to work (i.e. not lazy at all) and are always considerate and kind to others (i.e. never rude). The factor

skepticism combines the items worrying and nervousness. Approachability combines new ideas, talkative,

outgoing (i.e. not reserved) and stressed easily (i.e. not relaxed). Finally, Amiableness is a combination

of talkative and sociable but also forgiving and kind.

Table 2 shows the correlation between the Big Five factors and the factors derived from the conducted

PCA. The results suggest that our factors are relatively close to the Big Five factors. Our factor Cre-

ativeness is significantly correlated to the factor Openness from the Big Five model. Similarly, our factor

Skepticism can be clearly mapped to the factor Neuroticism, and, our factor Approachability to the Big

Five factor Extraversion. For the remaining two factors, Diligence and Amiableness we see correlations

with more than one factor or with none of the factors from the Big Five model. Overall, we conclude

that it is suitable to use the Big Five factors for our survey population as the results suggest a strong

correlation between our factors and the Big Five factors. The same validation technique has been followed

by (Rammstedt and John, 2007) to establish the equivalence of the BFI-S to the BFI-44.

Comparison of Sub-Sample Table 3 depicts the average score for each of the Big Five factors for

those individuals included in the TVSEP and the Add-on project. The results reveal that on average

the factors differ only slightly between the answers given in the TVSEP data and the Add-on project.

The factors Extraversion, Agreeableness and Neuroticism are not statistically different from each other.

Although, the factors Openness and Conscientiousness are statistically different from each other, the

10

Table 2: Correlation between Big Five and Factors from PCA

Openness Conscientiousness Extraversion Agreeableness Neuroticism

Creativeness 0.76 0.63 0.38 0.50 -0.23Diligence -0.37 0.51 0.12 0.64 -0.12Skepticism -0.02 0.13 -0.06 0.09 0.92Approachability 0.22 0.08 0.75 -0.24 0.11Amiableness -0.21 -0.42 0.46 0.32 0.07

Correlation higher than absolute 0.50 are shown in bold.Source: Authors’ calculations.

mean values are still very close together and do not contradict each other. Some of this variation might

also be the result of the different answer framing in the Add-on questionnaire. Due to this alteration

the answers are not 100 percent comparable. Moreover, questions were posed by enumerators and not

self-reported. This might have added some additional variation to the answers. The findings show that

the answers are consistent over time, which lets us to believe that overall the 15 survey questions were

posed in the correct way and that respondents understood them.

Table 3: Comparison of sample means

Mean TVSEP Mean Add-on Mean Difference

Openness 4.601 4.922 -0.321***Conscientiousness 5.549 5.743 -0.195***Extraversion 4.484 4.505 -0.021Agreeableness 5.593 5.589 0.004Neuroticism 3.399 3.264 0.135

Note: *, **, and *** denote significance at the 10, 5, and 1 percent levels for two-sided ttests.Source: Authors’ calculations.

Overall, the results from the Cronbach’s alpha and the PCA indicate that the personality factors in

our sample population are similar to the Big Five factors. Furthermore, the comparison between the

TVSEP data and the Add-on projects show that individuals answer consistent across the two surveys.

Thus, we conclude that the personality trait questions can be utilized to form the Big Five factors for

our study population. For comparability we use the average score of the original Big Five factors for the

remainder of the study.

4.2 Differences across countries and occupation categories

In this section we briefly review the differences across the Thai and Vietnamese sub sample. Furthermore,

we provide a descriptive overview of the different personality traits and some socio-economic character-

istics across the different occupation groups.

Table 4 reports the average scores on the question items for the Thai and Vietnamese sub sample.

The significance levels refer to the statistical difference between the two countries.

The results illustrate that on average, Vietnamese tend to report a higher score on all factors, except

Openness. County differences are highly significant for almost all of the 15 survey items. In general, the

Thai population appears to be more homogenous with respect to personality traits. This is not surprising

given the ethnic homogeneity in Thailand and the fact that Buddhism has a strong influence on all aspects

of live in Thailand. 97% of the Thailand sample population identify themselves as Thai and 99% follow

Buddhism. In contrast, rural Vietnam is more diverse in terms of both religion and ethnicity. 78% of the

sample population belongs to the majority ethnicity Kinh and 70% consider themselves as Atheists. The

results show that the rural Thai population scores higher in the questions related to Openness. Thus,they

11

are more artistic and have a more active imagination compared to the Vietnamese population. They also

score higher on efficiency and being talkative. In contrast, the Vietnamese are less reserved and less rude

but score significantly higher on the questions related to Neuroticism.

Table 4: Personality traits by country

Trait Variables Thailand Vietnam Diff.

Openness artistic 5.04 3.76 ***new ideas 4.59 4.46 *active imagination 4.37 3.97 ***

Conscientiousness work thorough 5.55 5.72 ***efficient 5.86 5.47 ***lazy r 5.61 6.23 ***

Extraversion talkative 4.98 4.57 ***sociable 5.06 4.99reserved r 3.39 4.10 ***

Agreeableness forgiving 5.80 5.81kind 5.96 5.88 *rude r 5.49 5.96 ***

Neuroticism worries 3.86 5.59 ***nervous 3.51 4.46 ***relaxed r 2.66 3.27 ***

N 1575 1595

Note: *, **, and *** denote significance at the 10, 5, and 1 percent levels for two-sided ttests.Source: Authors’ calculations.

Table 5 reports the average scores on the question items for the different categories of occupations –

subsistence farmers, commercial farmers, self-employed, casual workers, permanent workers, housewives

and non-workers – previously described. The significance levels refer to the statistical difference between

one particular group and subsistence farmers.

The results suggest that commercial farmers are significantly different from subsistence farmers along

the factors of Openness, Extraversion and Agreeableness as well as trust and patience. For example,

commercial farmers are less artistic than the average subsistence farmer as well as less reserved. They

also score higher on trust and patience.

Self-employed individuals differ along the factors of Openness and Conscientiousness. Self-employed

individuals are more open to new ideas and less lazy than the average subsistence farmer. They are also

better educated, with an average of 7.13 years of schooling. The group of casual workers differs along the

factors of Openness, Extraversion and Neuroticism. They have an active imagination and tend to worry

a lot as well as being nervous easily. They have a lower imagination, are less social and worry less.

In contrast to the casual worker, permanent workers differ along the factors of Conscientiousness and

Extraversion. They also have a higher internal locus of control. Permanently employed individuals are

significantly more efficient and sociable compared to the average subsistence farmer. With an average

age of 46.20 these individuals are significantly younger and with almost 10 years of education they are in

comparison the best educated group.

Housewives in our sample population score significantly different on Conscientiousness and Extraver-

sion. They are lazier and less sociable than the average farmer. Housewives score have a higher internal

locus of control and score lower on trust. Finally, the group of non-workers differs along the dimension

of Neuroticism. Interestingly, individuals who are not working worry less and are less nervous.

In sum, the results indicate that personality traits differ across Thailand and Vietnam as well as

across occupation groups. Comparing mean levels of personality traits at national levels is appropriate

and aids to understand the link between culture and psychology (Levine, 2001; Saucier Goldberg, 2001).

Overall, it reveals that countries and regions differ with regards to the Big Five. Different cultures,

12

political systems, religion, history etc. are also reflected in peoples personalities. Furthermore, the

comparison by occupation group indicates that individuals who opt for self-employment or permanent

wage-employment score higher along the factor of Conscientiousness. Furthermore, individuals in these

two groups are younger and significantly better educated compared to farmers.

13

Tab

le5:

Per

son

alit

ytr

aits

by

occ

up

atio

n

Trait

Varia

ble

sFarm

er

Farm

er

Self

-W

orker

Worker

Hou

se-

Not

sub

sist

en

ce

com

mercia

lem

plo

yed

casu

al

perm

an

ent

wif

ew

orkin

g

Op

enn

ess

Art

isti

c4.3

64.0

5***

4.4

94.3

94.6

04.6

84.8

0N

ewid

eas

4.5

84.4

94.8

2*

4.3

54.7

74.3

54.2

6A

ctiv

eim

agin

ati

on

4.2

34.0

7*

4.2

73.7

6**

4.3

73.9

84.0

7

Con

scie

nti

ou

s-W

ork

thoro

ugh

ly5.6

35.6

15.7

35.4

65.7

55.4

45.3

0n

ess

Effi

cien

t5.6

55.5

75.7

25.8

05.8

9**

5.5

75.4

1L

azy

2.1

02.0

01.8

5**

2.0

91.9

52.6

0**

2.3

1

Extr

aver

sion

Talk

ati

ve

4.7

44.7

94.9

34.7

24.8

64.8

14.3

0S

oci

ab

le5.0

64.9

84.9

64.7

2*

5.3

8*

4.6

0**

5.0

0R

eser

ved

4.3

33.9

8***

4.2

94.4

14.4

84.4

74.0

4

Agre

eab

le-

Forg

ivin

g5.8

55.7

35.7

95.6

85.9

25.8

15.8

5n

ess

Kin

d5.9

65.8

1**

5.9

25.7

5*

6.0

55.7

55.9

3R

ud

e2.3

32.2

02.2

22.4

22.3

12.3

72.3

7

Neu

roti

cism

Worr

ies

4.8

94.8

24.9

44.3

4***

4.8

54.6

54.0

9**

Ner

vou

s4.0

54.1

93.9

33.6

6*

3.7

94.3

63.2

0***

Rel

axed

5.0

24.9

65.0

45.0

45.2

04.9

65.0

2

Locu

sof

contr

ol

4.4

44.3

74.7

14.4

54.9

2*

5.2

8**

4.9

8T

rust

1.7

81.8

5*

1.7

31.6

81.7

61.6

1*

1.7

6R

isk

6.4

06.3

86.5

36.3

56.3

05.9

55.9

6P

ati

ence

6.3

56.6

4*

6.3

95.9

26.3

05.9

05.8

0

Age

(yea

rs)

50.1

748.8

7**

48.5

4***

48.0

7**

46.2

0***

48.1

449.3

1E

du

cati

on

(yea

rs)

6.2

56.4

07.1

3***

5.8

310.0

1***

5.8

96.7

2R

elig

iou

s(1

=re

ligio

us)

0.6

10.5

3***

0.6

00.8

6***

0.6

50.8

4***

0.8

0**

Eth

nic

ity

(1=

ma

jori

ty)

0.8

70.7

6***

0.9

7***

0.8

50.8

90.8

70.9

4G

end

er(1

=fe

male

)0.6

60.5

7***

0.7

00.6

10.5

3***

0.9

8***

0.5

6

N1134

672

328

162

210

110

54

Note

:*,

**,

an

d***

den

ote

sign

ifica

nce

at

the

10,

5,

an

d1

per

cent

level

sfo

rtw

o-s

ided

ttes

ts.

All

sub

gro

up

sare

test

edagain

stth

egro

up

’su

bsi

sten

cefa

rmer

’S

ou

rce:

Au

thors

’ca

lcu

lati

on

s.

14

4.3 Importance of personality traits for occupational attainment

Results from the multionominal logit regression are presented in Table 6. Using Subsistence farming as

the baseline occupation, the columns depict the relative odds ratio of one occupation category, each in

relation to being a farmer. Controls for individual, household and region-specific effects are included.

The pseudo R squared of 0.11 indicates that the model fit is acceptable.

The results show some heterogeneity across groups. For respondents, who score higher on Extraversion

the relative odds of choosing to take up Commercial Farming over Subsistence Farming increases. The

results match with the idea of an active and outgoing farmer. Moreover, higher levels of Conscientiousness

are associated with a higher likelihood of choosing Self-employment or Permanent wage employment over

Subsistence farming. Both are occupation categories that require someone to be well organized and

responsible. For the group of casual workers we see opting for this occupation category is not only

influenced by higher levels of conscientiousness, but also that lower levels of Openness and Patience

increase the odds of working in casual wage-employment. Interestingly, we see that having a higher

internal locus of control, i.e. a strong feeling that ones own life outcomes are contingent on their actions,

increases the relative odds of being a housewife. Furthermore, the results reveal a negative relationship

between Neuroticism and being unemployed. Meaning, that respondents, who are not working tend to

worry less and are less nervous, which seems puzzling, since unemployment is generally associated with

higher stress levels. Overall, our results suggest that personality traits, specifically, Conscientiousness,

are associated with the occupational attainment of individuals. In particular, for individuals who score

higher on Conscientiousness, the relative odds of choosing self-employment, permanent employment or

casual wage employment increase in relation to the odds of opting to become a Subsistence-Farmer. Our

results indicate that Conscientiousness is particularly important for choosing higher skilled jobs over

farming. This is crucial for development support policies as in order to generate better employment

opportunities in rural areas there is not only a need to develop suitable financial instruments, but also

to identify individuals who are willing to choose an occupation outside of the farming business.

The individual-level control variables indicate further important differences that relate to occupa-

tional attainment. The results confirm that for females the odds of being a housewife are substantially

larger than to engage in Subsistence-Farming. Furthermore, the odds of choosing self-employment over

Subsistence-Farming are also larger for females. This potentially hints at the fact that households di-

versify their income earning activities but keep farming one activity that is predominantly conducted by

males. Whereas women run small businesses from their home, which also gives them the opportunity to

take care of the children at the same time. Given, the rural setting of the sample, where women tend

to take care of the children, these results seem plausible. The results are also in line with the fact that

women are less likely to engage in Commercial Farming as well. The picture is different for individuals

who choose to take up permanent employment rather than becoming a Subsistence-Farmer. The odds

for choosing permanent employment over farming is significant but smaller for women meaning that in

our sample population it is rather men who opt for permanent wage employment. In terms of age the

results clearly show that the odds of choosing self-employment, working casually or permanently decrease

in relation to Subsistence-Farming with every year of age. This indicates that younger people opt for

occupations other than farming. Finally, for individuals with a higher level of education the odds ratio

of choosing self-employment or permanent employment are larger than choosing farming. Thus, higher

skilled individuals opt for more complex tasks that are likely related with higher income.

Overall, the results from our regression analysis suggest that Personality Traits are important factors

for the occupational attainment of individuals in rural Southeast Asia. Specifically, individuals who are

15

Table 6: Relation between personality traits and occupational attainment

CommercialFarming

Self-employed

Wage-employed(perma-nent)

Worker(casual)

Housewife NotWork-ing

(1) (2) (3) (4) (5) (6)

Openness -0.0668 0.0355 -0.0306 -0.213*** 0.0214 -0.0369(0.0433) (0.0566) (0.0704) (0.0741) (0.0926) (0.122)

Conscientiousness 0.0204 0.198** 0.267** 0.193* -0.159 -0.227(0.0622) (0.0838) (0.108) (0.105) (0.124) (0.155)

Extraversion 0.0948* -0.000387 -0.0105 -0.0533 -0.0859 -0.0278(0.0490) (0.0633) (0.0797) (0.0843) (0.103) (0.135)

Agreeableness -0.0817 -0.0467 0.0920 -0.0888 0.0713 0.156(0.0605) (0.0790) (0.102) (0.0998) (0.125) (0.172)

Neuroticism -0.00842 -0.00702 0.00163 0.0234 0.108 -0.280**(0.0489) (0.0622) (0.0793) (0.0835) (0.104) (0.135)

Locus of Controls -0.00435 0.0242 0.0302 -0.0300 0.0794** 0.0432(0.0186) (0.0233) (0.0291) (0.0321) (0.0375) (0.0504)

Trust 0.0622 -0.151 -0.0946 0.109 -0.0351 0.131(0.0765) (0.102) (0.123) (0.130) (0.166) (0.213)

Risk -0.00866 0.00969 -0.0500 -0.000409 -0.0327 -0.0540(0.0194) (0.0252) (0.0319) (0.0311) (0.0373) (0.0512)

Patience 0.0118 -0.0102 -0.0131 -0.0540* -0.0313 -0.0375(0.0191) (0.0240) (0.0304) (0.0296) (0.0366) (0.0503)

Age -0.00842 -0.0244*** -0.0310*** -0.0479*** -0.0220 -0.0322*(0.00635) (0.00825) (0.00924) (0.0109) (0.0135) (0.0172)

Years of Schooling 0.00786 0.0786*** 0.279*** -0.0324 -0.00273 0.0257(0.0174) (0.0212) (0.0240) (0.0306) (0.0367) (0.0480)

Religious -0.325** -0.313 0.0567 0.303 0.332 0.616(0.158) (0.205) (0.250) (0.336) (0.379) (0.534)

Ethnicity -0.170 1.655*** -0.209 -0.454 -0.339 -0.220(0.173) (0.366) (0.308) (0.317) (0.422) (0.721)

Gender -0.279** 0.272* -0.423** -0.357* 3.202*** -0.448(0.109) (0.148) (0.174) (0.186) (0.721) (0.296)

Constant -0.158 -3.066** -16.64 4.216*** -1.424 2.802(0.767) (1.210) (783.3) (1.349) (1.769) (2.112)

Controls x x x x x x

Observations 2,658 2,658 2,658 2,658 2,658 2,658R2 0.113 0.113 0.113 0.113 0.113 0.113Chi2 926.9 926.9 926.9 926.9 926.9 926.9

Note: *, **, and *** denote significance at the 10, 5, and 1 percent levels. Clustered standard errors in parentheses.Source: Authors’ calculations

16

more conscientious and are better educated opt for self-employment and permanent wage employment.

Furthermore, gender norms are quite persistent and women are more likely to take up housework or

engage in self-employment rather than farming.

5 Conclusion

In this paper, we analyze the relationship between individual personality traits and the occupational

attainment of individuals in emerging economies. To achieve this, we employ data from a comprehensive

household survey from Thailand and Vietnam collected under the Thailand Vietnam Socio Economic

Panel. As this is the first paper that uses the personality questions from this data set, our first research

question examines the validity of the Big Five model for our sample. The results from the Cronbach’s

alpha and the PCA confirm the validity of the survey measurement tool applied. An additional robustness

test is executed by comparing the results from one province with a sub sample that was interviewed under

an Add-on project. The answers of the respondents illustrate consistency over time.

In addition, we investigate the importance of personality traits across countries and different occupa-

tion groups. Our results suggest that on average personality traits are more homogeneous among the Thai

compared to the Vietnamese sub sample. While the Thai score higher on questions related to Openness,

the Vietnamese sample reports to be less reserved and less rude and scores higher on questions related

to Neuroticism. Furthermore, self-employed and permanently employed individuals are younger and bet-

ter educated and report higher levels of Openness, Conscientiousness and Extraversion. Interestingly,

individuals that are not working display lower levels of Neuroticism.

We further examine the influence of personality traits on occupational attainment and compare across

occupations that could be a possible alternative to subsistence farming in our sample. we perform a

multinomial logit estimation and find that the Big Five and Conscientiousness in particular are the most

important predictor of occupational attainment of an individual in Southeast Asia. Other non-cognitive

skills, such as locus of control, trust, risk and patience play an inferior role in predicting occupational at-

tainment. Conscientious and more educated individuals opt for self-employment or permanent wage jobs.

Women are mostly engaged as housewives. However, women are more likely to prefer self-employment

over farming.

Our study contributes to the literature by providing empirical evidence not only in the context of

developing countries but also in reference to rural job markets. The results emphasize the important role

of personality traits in individual decision making. From a policy perspective, a better understanding of

personality traits would aid in efficient policy making. The need to rethink development policy to account

for human factors has been widely identified (World Bank, 2015). Success of most development policies

is contingent on an individual’s participation, which again depends on the individual’s personality. This

is especially important for labor market policies. While policy makers can improve employment services

and incentivize self-employment through offering micro grants, it is up to the population to size these

opportunities.

In the next step, we aim to utilize the panel structure of TVSEP and look at the family background

of the respondents and their past shock experience.

17

References

Attanasio, Orazio, Sarah Cattan, Emla Fitzsimons, Costas Meghir, and Marta Rubio-Codina, “Estimating

the Production Function for Human Capital: Results from a Randomized Control Trial in Colombial.” NBER Working

Paper No. 20965, National Bureau of Economic Research, Cambridge, MA, USA 2015.

Backhaus, Klaus, Bernd Erichson, Wulff Plinke, and Rolf Weiber, Multivariate Analysemethoden: Eine anwen-

dungsorientierte Einfhrung, 13th ed., Berlin, Heidelberg, New York: Springer-Verlag, 2011.

Banerjee, Abhijit V and Esther Duflo, 2007, “The Economic Lives of the Poor.” The journal of economic perspectives

: a journal of the American Economic Association, 21 (1), 141–167.

Barrick, Murray R. and Michael K. Mount, 1991, “The Big Five Personality Dimensions and Job Performance: A

Meta Analysis.” Personal Psychology, 44 (1), 1–26.

Becker, Anke, Thomas Deckers, Thomas Dohmen, Armin Falk, and Fabian Kosse, 2012, “The Relationship

Between Economic Preferences and Psychological Personality Measures.” Annual Review of Economics, 4 (1), 453–478.

Ben-Porath, Yoram, 1967, “The Production of Human Capital and the Life Cycle of Earnings.” Journal of Political

Economy, 75 (4), 352–365.

Ben-Porath, Yoram, “The Production of Human Capital over Time.” in W. Lee Hansen, ed., Education, Income and

Human Capital, New York, U.S.: National Bureau of Economic Research, 1970, pp. 129–147.

Bentolila, Samuel, Claudio Michelacci, and Javier Suarez, 2010, “Social Contacts and Occupational Choice.”

Economica, 77 (305), 20–45.

Blau, Peter M., John W. Gustad, Richard Jessor, Herbert S. Parnes, and Richard C. Wilcoc, 1956, “Occu-

pational Choice: A Conceptual Framework.” Industrial and Labor Relations Review, 9 (4), 531–543.

Caliendo, Marco, Deborah A. Cobb-Clark, and Arne Uhlendorff, 2015, “Locus of Control and Job Search Strate-

gies.” Review of Economics and Statistics, 97 (1), 88–103.

Caliendo, Marco, Frank Fossen, and Alexander Kritikos, 2010, “The impact of risk attitudes on entrepreneurial

survival.” Journal of Economic Behavior & Organization, 76 (1), 45–63.

Caliendo, Marco, Frank Fossen, and Alexander Kritikos, 2012, “Trust, positive reciprocity, and negative reciprocity:

Do these traits impact entrepreneurial dynamics?” Journal of Economic Psychology, 33 (2), 394–409.

Caliendo, Marco, Frank Fossen, and Alexander S. Kritikos, 2014, “Personality characteristics and the decisions to

become and stay self-employed.” Small Business Economics, 42 (4), 787–814.

Campbell, Duncan, The global crisis: causes, responses and challenges, Geneva: International Labour Office, 2011.

Campbell, Duncan and Ishraq Rayeed Ahmed, “The Labour Market in Developing Countries.” in “in” 2012.

Cawley, John, James Heckman, and Edward Vytlacil, 2001, “Three observations on wages and measured cognitive

ability.” Labour Economics, 8 (4), 419 – 442.

Cobb-Clark, Deborah A. and Michelle Tan, 2011, “Non-cognitive skills, occupational attainment, and relative wages.”

Labour Economics, 18, 1–13.

Costa, Paul T. Jr. and Robert R. McCrae, “Revised NEO Personality Inventory (NEO-PI-R) and NEO Five Factor

Inventory (NEO-FFI) Professional Manual.” Technical Report, Odessa, FL, USA 1992.

Costa, Paul T. Jr. and Robert R. McCrae, 1997, “Personality trait structure as a human universal.” American

Psychologist, 52, 587–596.

Cronbach, Lee J., 1951, “Coefficient alpha and the internal structure of tests.” Psychometrika, 16 (3), 297–334.

Cunha, Flavio, James J. Heckman, Lance J. Lochner, and Dimitriy V. Masterov, “Interpreting the Evidence

on Life Cycle Skill Formation.” in Eric A. Hanusheck and Finis Welch, eds., Handbook of the Economics of Education

(Vol. 1), Amsterdam and Oxford: Elsevier, 2006, pp. 697–805.

18

Dohmen, Thomas, Armin Falk, David Huffman, and Uwe Sunde, 2009, “Homo Reciprocans: Survey Evidence on

Behavioural Outcomes*.” The Economic Journal, 119 (536), 592–612.

Dustmann, Christian, Francesco Fasani, Xin Meng, and Luigi Minale, 2017, “Risk Attitudes and Household

Migration Decisions.” SSRN Electronic Journal.

Erdheim, Jesse, Mo Wang, and Michael J. Zickar, 2006, “Linking the Big Five personality constructs to organizational

commitment.” Personality and Individual Differences, 41 (5), 959–970.

Falk, Armin, Anke Becker, Thomas Dohmen, Benjamin Enke, David Huffman, and Uwe Sunde, “The

preference survey module: A validated instrument for measuring risk, time, and social preferencesl.” IZA Discussion

Paper No. 9674, Institute of Labor Economics, Bonn, Germany 2016.

Falk, Armin, Anke Becker, Thomas Dohmen, Benjamin Enke, David Huffman, and Uwe Sunde, 2018, “Global

Evidence on Economic Preferences.” The Quarterly Journal of Economics, 133 (4), 1645–1692.

Finnie, Ross and Ronald Meng, 2002, “Minorities, Cognitive Skills and Incomes of Canadians.” Canadian Public Policy

/ Analyse de Politiques, 28 (2), 257–273.

Ford, J. Kevin, Robert C. MacCallum, and Marianne Tait, 1986, “The application of exploratory factor analysis

in applied psychology: A critical review and analysis.” Personnel Psychology, 39 (2), 291–314.

Gertler, Paul, James Heckman, Rodrigo Pinto, Arianna Zanolini, Christel Vermeersch, Susan Walker,

Susan M. Chang, and Sally Grantham-McGregor, 2014, “Labor market returns to an early childhood stimulation

intervention in Jamaica.” Science, 344, 998–1001.

Gloede, Oliver, Lukas Menkoff, and Hermann Waibel, 2015, “Shocks, Individual Risk Attitude, and Vulnerability

to Poverty among Rural Households in Thailand and Vietnam.” World Development, 71, 55–78.

Gollin, Douglas, David Lagakos, and Michael E. Waugh, 2014, “The Agricultural Productivity Gap.” Quarterly

Journal of Economics, 129 (2), 939–993.

Guerra, N., K. Modecki, and W. Cunningham, “Developing Socio-Emotional Skills for the Labor Market - The

Practise Model.” Policy Research Working Paper No. 7123, World Bank, Washington D.C., USA 2016.

Hair, Joseph F., William C. Black, Barry J. Babin, and Rolph E. Anderson, Multivariate data analysis, 7 ed.,

Harlow, Essex, UK: Pearson Education Limited, 2009.

Hanushek, Eric A. and Ludger Woessmann, 2008, “The Role of Cognitive Skills in Economic Development.” Journal

of Economic Literature, 46 (3), 607–68.

Hardeweg, Bernd, Lukas Menkhoff, and Hermann Waibel, 2013, “Experimentally Validated Survey Evidence on

Individual Risk Attitudes in Rural Thailand.” Economic Development and Cultural Change, 61 (4), 859–888.

Hardeweg, Bernd, Stephan Klasen, and Hermann Waibel, “Establishing a Database for Vulnerability Assessment.”

in Stephan Klasen and Hermann Waibel, eds., Vulnerability to Poverty: Theory, Measurements and Determinants with

Case Studies from Thailand and Vietnam, Basingstoke, UK: Palgrave Macmillan, 2013, pp. 50–79.

Heckman, James and Flavio Cunha, 2007, “The Technology of Skill Formation.” American Economic Review, 97 (2),

31–47.

Heckman, James J., 1976, “A Life-Cycle Model of Earnings, Learning, and Consumption.” Journal of Political Economy,

84 (4, Part 2), S9–S44.

Heineck, Guido and Silke Anger, 2010, “The returns to cognitive abilities and personality traits in Germany.” Labour

Economics, 17 (3), 535–546.

Hofstede, Geert and Robert R. McCrae, 2004, “Personality and Culture Revisited: Linking Traits and Dimensions

of Culture.” Cross-Cultural Research, 38 (1), 52–88.

Jencks, Christopher, Susan Bartlett Mary Corcoran James Crouse David Eaglesfield Gregory Jackson Kent

McCelland Peter Mueser Michael Olneck Joseph Schwartz Sherry Ward and Jill Williams, Who gets ahead?

The Determinants of Economic Success in America, New York, U.S.: Basic Books, 1979.

19

John, Katrin and Stephan L. Thomsen, 2014, “Heterogeneous returns to personality: the role of occupational choice.”

Empirical Economics, 47, 553–592.

Judge, Timothy A., Daniel Heller, and Michael K. Mount, 2002, “Five-factor Model of Personality and Job

Satisfaction: A Meta-analysis.” Journal of Applied Psychology, 87, 530–41.

Judge, Timothy A., Joyce E. Bono, Remus Ilies, and Megan W. Gerhardt, 2002, “Personality and Leadership:

A Qualitative and Quantitative Review.” Journal of Applied Psychology, 87, 765–80.

Kluhs, Theres, Melanie Koch, and Wiebke Stein, “Don’t Expect Too Much: The Effect of Biased Expectations on

(Over)-Indebtedness.” 2019. Mimeo.

Laajaj, R. and K. Macours, “Measuring SKills in Developing Countries.” Policy Research Working Paper No. 8000,

World Bank, Washington D.C., USA 2017.

McCrae, Robert R. and Oliver P. John, 1992, “An Introduction to the Five-Factor Model and Its Applications.”

Journal of Personality, 60 (2), 175–215.

McIntosh, Steven and Anna Vignoles, 2001, “Measuring and Assessing the Impact of Basic Skills on Labour Market

Outcomes.” Oxford Economic Papers, 53 (3), 453–481.

Mincer, Jacob, 1970, “The Distribution of Labor Incomes: A Survey with Special Reference to the Human Capital

Approach.” Journal of Economic Literature, 8 (1), 1–26.

Mincer, Jacob, “Schooling, Experience, and Earnings.” Technical Report, National Bureau of Economic Research, New

York 1974.

Mueller, Gerrit and Erik Plug, 2006, “Estimating the effect of personality on male and female earnings.” Industrial

and Labor Relations Review, 60 (1), 3–22.

Nyhus, Ellen K. and Empar Pons, 2005, “The effects of personality on earnings.” Journal of Economic Psychology,

26, 363–384.

Obschonka, Martin and Michael Stuetzer, 2017, “Integrating psychological approaches to entrepreneurship: the

Entrepreneurial Personality System (EPS).” Small Bus Econ, 49, 203–231.

OECD, “Working age population.” https://data.oecd.org/pop/working-age-population.htm 2019. [Online; accessed

12-June-2019].

Ones, Deniz S., Chockalingam Viswesvaran, and Frank L. Schmidt, 2003, “Personality and absenteeism: a meta-

analysis of integrity tests.” European Journal of Personality, 17 (S1), S19–S38.

Osborne Groves, Melissa, 2003, “How important is your personality? Labor market returns to personality for women

in the US and UK.” Journal of Economic Psychology, 26, 827–841.

Parvathi, Priyanka and Trung Thanh Nguyen, 2018, “Is Environmental Income Reporting Evasive in Household

Surveys? Evidence From Rural Poor in Laos.” Ecological Economics, 143, 218–226.

Piatek, Remi and Pia Pinger, “Maintaining (Locus of) Control? Assessing the Impact of Locus of Control on Education

Decisions and Wages.” SOEP Papers No. 338, German Institute for Economic Research (DIW), Berlin, Germany 2010.

Piatek, Rmi and Pia Pinger, 2016, “Maintaining (Locus of) Control? Data Combination for the Identification and

Inference of Factor Structure Models.” Journal of Applied Econometrics, 31 (4), 734–755.

Rammstedt, Beatrice, 2007, “The 10-Item Big Five Inventory: Norm Values and Investigation of Sociodemographic

Effects Based on a German Population Representative Sample.” European Journal of Psychological Assessment, 23 (3),

193–201.

Rammstedt, Beatrice and Oliver P. John, 2007, “Measuring personality in one minute or less: A 10-item short version

of the Big Five Inventory in English and German.” Journal of Research in Personality, 41 (1), 203–212.

Rolland, Jean-Pierre, The Cross-Cultural Generalizability of the Five-Factor Model of Personality, Boston, MA: Springer

US,

20

Rotter, Julian B., 1966, “Generalized expectancies for internal versus external control of reinforcement.” Psychological

Monographs: General and Applied, 80 (1), 1–28.

Schafer, Konrad C., “The Influence of Personality Traits on Private Retirement Savings in Germany.” SOEP Papers No.

867, German Institute for Economic Research (DIW), Berlin, Germany 2016.

Schmidt, Peter and Robert P. Strauss, 1975, “The Prediction of Occupation Using Multiple Logit Models.” Interna-

tional Economic Review, 16 (2), 471–486.

Schmitt, David P., Ano Realo, Martin Voracek, and Juri Allik, 2008, “Why Cant a Man Be More Like a Woman?

Sex Differences in Big Five Personality Traits Across 55 Cultures.” Journal of Personality and Social Psychology, 94 (1),

168–182.

Schmitt, David P., Juri Allik, Robert R. McCrae, and Veronica Benet-Martnez, 2007, “The Geographic Dis-

tribution of Big Five Personality Traits: Patterns and Profiles of Human Self-Description Across 56 Nations.” Journal

of Cross-Cultural Psychology, 38 (2), 173–212.

Schnitzlein, Daniel and Jens Stephani, 2016, “Locus of Control and low-wage mobility.”

Seibert, Scott E. and Maria L. Kraimer, 2001, “The Five-Factor Model of Personality and Career Success.” Journal

of Vocational Behavior, 58, 1–21.

Sharma, Rasadhika, Tung Nguyen, Ulrike Grote, and Trung Thanh Nguyen, “Changing livelihoods in rural

Cambodia: Evidence from panel household data in Stung Treng. Center of Development Research.” Working Paper 149,

ZEF - Centre for Development Research, Bonn, Germany 2016.

Sharma, Smriti and Finn Tarp, 2018, “Does managerial personality matter? Evidence from firms in Vietnam.” Journal

of Economic Behavior & Organization, 150, 432–445.

Sohns, Franziska and Javier Revilla Diez, 2016, “Self-Employment and its influence of the vulnerability to poverty of

households in rural Vietnam a panel data analysis.” Geographical Review, 107 (2), 336–359.

Stuetzer, Michael, David B. Audretsch, , Martin Obschonka, Samuel D. Gosling, Peter J. Rentfrow, and

Jeff Potter, 2018, “Entrepreneurship culture, knowledge spillovers and the growth of regions.” Regional Studies, 52 (5),

608–618.

UNHCS, “General Assembly, Special Session for an overall review appraisal of the implementation of the Habitat Agenda.

Press Kit. Urbanization: Facts and Figures.” https://www.un.org/ga/Istanbul+5/bg10.htm 2001. [Online; accessed 12-

June-2019].

Wells, Robert, Roger Ham, and P. N. (Raja) Junankar, 2016, “An examination of personality in occupational

outcomes: antagonistic managers, careless workers and extraverted salespeople.” Applied Economics, 48 (7), 636–651.

Wichert, Laura and Winfried Pohlmeier, “Female Labor Force Participation and the Big Five.” Discussion Paper No.

10-003, ZEW - Centre for European Economic Research, Berlin, Germany 2010.

World Bank, “World Development Report 2015: Mind, Society, and Behavior.” Technical Report, World Bank, Washing-

ton, D.C., U.S. 2015.

World Bank, Riding the Wave: An East Asian Miracle for the 21st Century, Washington, D.C., USA: World Bank, 2018.

Yomaboot, Panida and Andrew J. Cooper, 2016, “Factor Structure and Psychometric Properties of the International

Personality Item Pool-NEO (IPIP-NEO) Thai Version.” Journal of Somdet Chaopraya Institute of Psychiatry, 10 (5),

36–49.

21

A Appendix

Table A.1: Example of adjectives defining the Big Five factors

Factor Facets/Adjectives

Extraversion Active, Assertive, Energetic, Enthusiastic, Outgoing, Talkative

Agreeableness Appreciative, Forgiving, Generous, Kind, Sympathetic, Trusting

Conscientiousness Efficient, Organized, Planful, Reliable, Responsible, Thorough

Neuroticism Anxious, Self-Pitying, Tense, Touchy, Unstable, Worrying

Openness Artistic, Curious, Imaginative, Insightful, Original, Wide interests

Figure A.1: Overview of Survey Region

22

Table A.2: Overview of survey questions

Do you see yourself as someone who.

is sometimes a bit rude to others?works thoroughly?is talkative?worries a lot?is original, comes up with new ideas?has a forgiving nature?tends to be lazy?is outgoing, sociable?gets nervous easily?values artistic, aesthetic experiences?is considerate and kind to almost everyone?does tasks efficiently?is reserved?is relaxed, handles stress well?has an active imagination?

Figure A.2: Scree plot of eigenvalues after PCA

Table A.3: Cronbach’s Alpha

Personality Trait Cronbach’s alpha No. of items

Openness 0.60 3Conscientiousness 0.55 3Extraversion 0.42 3Agreeableness 0.58 3Neuroticism 0.56 3

All Traits 0.67 15

23

Table A.4: Factor Loadings according to PCA

BFI-Items Factor 1 Factor 2 Factor 3 Factor 4 Factor 5Creativeness Diligence Skepticism Approachability Amiableness

Artistic 0,30 -0,27 -0,10 0,01 0,08New Ideas 0,31 -0,12 0,12 0,31 -0,35Active Imagnation 0,32 -0,26 0,10 0,05 -0,14Work thoroughly 0,30 0,22 0,10 0,04 -0,39Efficient 0,35 0,11 -0,08 -0,06 -0,30Lazy (reversed) 0,10 0,53 -0,08 0,05 -0,31Talkative 0,24 -0,12 -0,03 0,45 0,22Sociable 0,32 -0,02 0,00 0,30 0,30Reserved (reversed) -0,15 0,24 -0,12 0,65 0,15Forgiving 0,28 0,25 0,04 -0,19 0,48Kind 0,35 0,23 0,00 -0,18 0,33Rude (reversed) 0,00 0,53 -0,14 -0,04 0,06Worries 0,00 0,15 0,67 0,01 -0,01Nervous 0,00 0,01 0,66 -0,02 0,12Relaxed -0,31 0,12 0,17 0,32 0,00

24